SuperAPI AI Directory

revenue-forecasting

General & other

View repository

Updated 10-15 16:51

About this skill

Revenue Forecasting

Stripe MCP Setup

Requires Stripe MCP (https://docs.stripe.com/mcp). GetMcpTools before CallMcpTool. Paginate all lists (limit: 100, starting_after while has_more). Amounts in cents. Read-only for audits; confirm before writes.

Inputs to gather

  1. Current MRR (via mrr-arr-snapshot logic)
  2. New subs / month — count active subs created in last 30/60/90d (created filter)
  3. Monthly churn % (via churn-analysis)
  4. Avg new MRR per sub

Model (simple)

Next month MRR = current_MRR * (1 - churn_rate) + new_MRR_added

Project 3/6/12 months. Show best / base / worst (churn +/-2pp, growth +/-20%).

Output Template

Revenue Forecast (base case)

Current MRR: $[X]  |  Monthly growth: [+X%]  |  Churn: [X%]

| Month | MRR (worst) | MRR (base) | MRR (best) |
|-------|-------------|------------|------------|
| +3 | | | |
| +6 | | | |
| +12 | | | |

ARR run-rate (12mo base): $[X]
Key assumption risks: [list]

State all assumptions explicitly. Forecast quality degrades past 6 months.